THE GRAND RAPIDS STUDY
The city selected as the "fluoride-free" city for comparison with Grand Rapids was Muskegon, Michigan, "whose source of drinking water supply and geographical and climatological characteristics were similar to those of Grand Rapids" (Dean et al. 1950). This city was the only control one in which the caries attack rates in each year were published for each yearly age group. Unfortunately, its usefulness was marred by a number of features:
Large differences in sample size. The reliability of a mean rate is greatly influenced by the number of observations on which it is based. Because of the small number of subjects included in some age groups in some years in Muskegon, little reliance can be placed on the values stated. In twelve categories fewer than twenty children were examined. One "group" consisted of only one child, whereas one contained 462 children (Arnold et al., 1953). In the test city the variation in sample size was even greater, from 18,606 to 3 subjects.
Sampling by school class. "selected age groups of children are examined within each of the schools. Selection is made on the basis of school grade or class, using all children present in a class or grade of a school." (Arnold et al. 1953). These grades or classes were examined in 1945, an additional grade being examined in 1946, 1947 and 1949, and two more in 1950, making a total of eight grades in 1950 and 1951. In the last mentioned year Muskegon ceased to act as a control.
Different methods of sampling. In Grand Rapids the "annual study sample was selected after careful review of census data and consultation with city planning department officials. On the basis of available information, the 31 school districts of Grand Rapids were classified on a socio-economic basis. From the 79 schools in these districts, 25 representative schools were selected, and the examiners were assigned schools on a basis of equal sized samples of comparable population groups" (Arnold, et al., 1953). However, that strange procedure was not followed in Muskegon, the same authors stating that "In Muskegon, the annual examinations have been conducted in almost all schools, excluding only a few small schools on the periphery of the city where many students are from rural areas."
Changes in examiners. In the report of this study up to and including 1951 (in July 1951 Muskegon ceased to be a control), it was stated: "There have been changes in dental examiners with the exception of one officer who has participated in each series of examinations. Each new examiner has been calibrated against this one officer to standardize diagnostic criteria" (Arnold et al., 1953). The degree of success of this odd calibration procedure was not stated.
Examiner variability not assessed. In 1953 Arnold et al said that "Bite-wing X-ray examinations were made of a representative sample of children examined by the different examiners to evaluate, in part, the "examiner error"." However, such a procedure cannot replace the data that could have been obtained by a correctly designed examination process, which would have enabled the determination of between-examiner and within-examiner variability.
Late examination of control city. The authors of this study did not determine the caries attack rates in children in Muskegon prior to accepting this city as a suitable control, for comparison of two of their statements makes it clear that the results of the basic examination in the control city were not known until after the water of the test city was fluoridated. They stated that "Fluoridation of the Grand Rapids water supply was started January 25, 1945" (Dean et al., 1950), and that the "basic examinations in Muskegon were not done until late spring of 1945" (Arnold et al., 1953).
Water of control fluoridated. Another fact which limits the usefulness of Muskegon as a control city is that its water supply was fluoridated in July 1951 (Arnold et al., 1953), so that the results obtained after that date had no value as controls for those of Grand Rapids. This event occurred six and a half years after the institution of fluoridation in Grand Rapids, and therefore at a time when, in the latter city, few of the permanent teeth had erupted in the children that had been ingesting fluoridated water since birth.
Ignorance of commencement of fluoridation in control city. The fact that Muskegon had ceased to be a control by having its water fluoridated in July 1951 was not always realized. For instance, Black (1955) in a paper "Presented before Section on Public Health Dentistry, ninety-fifth annual session, American Dental Association, Miami, Fla., November 8, 1954" - over three years after the institution of fluoridation in Muskegon (Arnold et al. 1953) - said that ,i>"At Muskegon, Mich, the control city where fluoride-free water is used, the incidence of dental caries is unchanged and approximates the norm." Black was commenting on the findings made "After eight years of fluoridation at Grand Rapids" (that is eighteen months after the fluoridation of the Muskegon water). No information has been found in the literature with regard to the "national norm," in fact a feature of these trials has been the divergent pre-fluoridation rates. The differing caries attack rates seen in different localities, even in the same state, were illustrated by Hagan (1947) and by Hadjimarkos and Storvick (1949, 1950).
A similar statement to that of Black (1955) was made by Martin (1956) in a lecture delivered at the fourteenth Congress, Australian Dental Association, Melbourne, March 1956; that is, over four and a half years after the Muskegon water was fluoridated. He said: "The decay rates in the fluoride free control area (Muskegon) have remained unchanged." The paragraph containing the above-mentioned quotation cites as reference Arnold et al. (1953), who in that paper said: "The water supply at Muskegon remained unchanged until July 1951, when the city started adding fluorides to its water supply."
Variations in Muskegon rates. Tables 3 and 4 of the report of the seventh year of the Grand Rapids study (Arnold et al., 1953) show that both the def (decayed, extraction indicated, or filled deciduous teeth) and the DMF rates reported from Muskegon from year to year differed considerably from those of the first examination. Despite this fact, the statement that the incidence of dental caries in Muskegon was unchanged was made by Black (1955) and Martin (1956), amongst others. These statements are at variance with that of the authors of the study (Arnold et al., 1953), for they mentioned the changes in these words: "A similar comparison of results at Muskegon shows the percentage reduction to range from 1.5 percent in 6-year olds to a high of 15.5 percent in the 11 year olds" in the permanent teeth.
The magnitude of the changes. The statement which has been quoted above does not reveal the magnitude of the changes which were observed in the DMF rates in the control city. The percentage reductions given were obtained by the method commonly used in all these studies, that is, the difference between the most recent and the original DMF rate was expressed as a percentage of the original rate, the variations obtained in the intervening years being ignored. The changes which occurred would have been more obvious if, for instance, the results for Muskegon had been computed in 1946 instead of in 1951. In that case the "reduction" would have been 40.7 per cent instead of 1.5 per cent in the six-year old, and 32.7 per cent instead of 15.5 per cent in the eleven-year-old children.
Sampling error. Arnold et al. (1953) stated that the percentage reductions obtained in Muskegon "may, in part, represent sampling error." It was not conceded that such an error could also apply to the results from the test city, nor were suggestions made as to the cause of the remainder of the reductions in the control city. These cannot be attributed to changes in the water supply, for they stated: "The water supply at Muskegon remained unchanged until July 1951, when this city started adding fluorides to its water supply" (Arnold et al., 1953).
Variation in Muskegon. The variability from year to year in the mean rates reported from Muskegon, which, it will be recalled, were stated to have remained unchanged (Black, 1955; Martin, 1956) are illustrated in Figs. 1 and 2. The data from which these figures were drawn are from Tables 2 and 3 (with errata corrected) of Arnold, Dean, Jay and Knutson (1956). The point shown in Fig. 2 for the sixteen year-old children in 1946 should be disregarded, as this age "group" consisted of only one child.
Comparison of the series of baseline rates for DMF in Grand Rapids and in Muskegon does not reveal that one series was consistently higher than the other. However, with the exception of the eleven and twelve-year old children, the def rates in Muskegon were higher than they were in the test city.
Comparison with Aurora. Arnold et al. (1953) said: "To establish what might be termed an "expectancy curve" for this study, a natural fluoride area, the city of Aurora, Ill., was selected. The Aurora water supply contains 1.2 ppm F and has a reliable "history of constancy back to 1895." It was not stated whether factors other than the fluoride content of the water supply were considered in selecting this city. The fact that other influences can be of importance was shown in the recent study by Russell (1956) in Montgomery-Prince Georges counties. Prior to the institution of fluoridation, in the total sample of subjects the def rates for children whose mean ages were 5.44, 6.47, 7.45 and 8.49 years were lower than those of children of similar ages in Aurora.
Limitations of Aurora data. The caries attack rates reported from Aurora consist of a single series obtained by several examiners in 1945-6. Therefore, there is no information with regard to variations from year to year in the mean value of the rates, and examiner variability was not considered.
The mean number of def deciduous teeth per child in Muskegon, Michigan, the "fluoride-free" control city for Grand Rapids, Michigan, at each year of examination. The 1945 examination was made in the "late spring", those of the other years, in October and November. Data from Table 2, Arnold et al., 1956. Three months prior to the 1951 examination the water of this city was fluoridated. It has been stated (see p. 145 - 146) that these rates "remained unchanged" during the period shown.
THE GRAND RAPIDS STUDY
The mean number of DMF permanent teeth per child in Muskegon, Michigan, the "fluoride-free" control city, at each year of examination. Data from Table 3, Arnold, et al., 1956. It has been stated (see p. 145 - 146) that these rates "remained unchanged" during the period shown.